IMPROVED ALUMINA SOL FCC CATALYSTS MEET CHALLENGES OF THE 1990'S

Ulrich Alkemade, Steven Cartlidge, John M. Thompson
Grace GmbH
Worms, West Germany

Improved alumina sol, fluid catalytic cracking (FCC) catalysts allow refiners to upgrade heavier feeds containing high amounts of vanadium and nickel to give premium octane gasolines.

When an alumina sol binder is used in catalyst preparation, the desired nickel and vanadium-tolerant alumina phase is formed by precise control of the catalyst finishing conditions.

New zeolite formulas also increase gasoline motor octane number and lower gasoline octane sensitivity maintaining gasoline yields.

Innovations in FCC catalyst design are meeting the technical requirements of an increasingly demanding petroleum refining industry. Major advances in catalyst manufacturing provide a means for the rapid transfer of innovations from the research laboratory to a precise control environment essential for high-quality FCC catalyst production.

NICKEL-TOLERANT MATRIX

The profit margin of an FCCU is often constrained by the wet-gas compressor or the coke burning capacity of the unit. The wet-gas compressor limitations are more critical when nickel on the FCC catalyst matrix actively dehydrogenates the FCCU feed.

Unfortunately, nickel is an excellent dehydrogenation catalyst which converts hydrocarbons to coke and hydrogen.

Chemical passivation of nickel, by adding antimony compounds to the fresh feed, is practiced by many refineries. To reduce the dehydrogenation activity of metallic nickel, the FCC catalyst can be reacted with hydrogen sulfide and steam premixed into the feed to produce less reactive nickel sulfides and oxides.

Chemical passivation of nickel by interaction with the FCC catalyst matrix surface can also take place. For example, silica-rich FCC catalysts have been proposed to inhibit nickel activity. 1-4

Particulate alumina added to FCC catalysts will react with nickel during catalyst regeneration to form nickel aluminate spinal structures. 5 In addition to chemical passivation, the dehydrogenation activity of a particular nickel compound is related to the dispersion or surface area of the nickel species on the catalyst matrix. Nickel from the feed becomes occluded within the surface of the catalyst matrix.

If each nickel atom is separately anchored onto the matrix the number of sites available for catalytic dehydrogenation will be at a maximum.

A nickel-tolerant catalyst matrix must promote the selfpassivation of nickel by allowing nickel to form large agglomerates having low surface areas.6 Particulate alumina has been shown to act as a sink for nickel on FCC catalysts .7

The alumina chemistry, crystal size, and accessibility within the FCC catalyst particle greatly affect the reactivity towards nickel. When an alumina sol binder is used in catalyst preparation, the desired nickel-tolerant alumina phase is formed by precise control of the finishing conditions.

New alumina sol catalysts, such as Super Resoc, give low yields of hydrogen and coke in the presence of nickel. Catalyst steamed matrix surface area is one parameter which characterizes the systematic improvements to the alumina sol matrix.

The systematic decrease of hydrogen and coke yields for nickel impregnated catalysts (Fig. 1) indicates the flexibility in matrix design.

Commercial data presented in Fig. 2 were collected from a European FCCU operation where the equilibrium catalyst contained up to 4,500 ppm Ni + Y. The expected decrease in hydrogen-to-methane ratio during the inventory change over to the improved alumina sol matrix could be readily monitored.

The low hydrogen and coke yields obtained for the catalyst were used to increase the unit operational severity leading to maximum conversion, minimum slurry oil production, and improved gasoline octane production, especially motor octane number (MON).

It should be noted that the chemical composition of the matrix remains constant, but the catalyst's tolerance to nickel poisons increases. The alumina content of the improved alumina sol matrix catalyst is about 41 wt %, and the good bottoms cracking activity of the alumina sol matrix is retained.

An improved distribution of acid sites on the matrix leads to low coke yields which allows bottoms cracking to be further exploited in an FCCU operation by increasing the cat-to-oil severity.

SILICA-RICH ZEOLITES

Catalyst deactivation by vanadium poisoning has been said to be related to the attack of the zeolite by vanadic acid.8 Improvements in catalyst vanadium tolerance are obtained by protecting the zeolite using matrix "vanadium getters" (traps) and by improving the intrinsic zeolite stability.

Vanadium porphyrins are thermally decomposed in the FCCU riser, and vanadium is carried over on the spent FCC catalyst into the regenerator.

A vanadium-tolerant FCC catalyst must contain zeolites which have exceptionally good hydrothermal stability and activity retention in the presence of vanadium. A key parameter which determines catalyst activity retention (Fig. 3) is the silica-to-alumina ratio of the zeolite framework.

The proportion of silica in the framework of zeolite Y increases from the starting soda-Y to the product ultrastable Y. This is the Z-14US ultra-stable zeolite.9

Research and development programs using experimental design techniques, coupled with modern manufacturing equipment, have been carried out. The results show that three variables, starting soda-Y silica-to-alumina ratio, degree of sodium exchange, and hydrothermal calcination conditions, determine ultra-stable Y stability and activity.

All three variables are highly significant factors which, when simultaneously optimized, considerably improve the zeolite surface area and activity retention after severe steam treatments representative of FCCU regenerator conditions.

Increases in ultra-stable Y hydrothermal activity (Fig. 4) are being obtained as part of a continual R&D commitment to improve catalyst activities and cracking selectivities under the most severe regenerator environments. The improved zeolite stability is obtained along with the mesoporous channel system characteristic of USY zeolites.

The wide mesopores (about 40 A dia.) in USY allow the access of high-boiling feed hydrocarbons into the zeolite, thereby enhancing bottom-of-the-barrel conversion.

VANADIUM TOLERANCE TESTS

It is important to relate the laboratory evaluation of catalyst metals tolerance to that in actual FCC operations. The most important difference between laboratory tests and FCCU metals tolerance evaluations lies in the method used to load the metals onto the catalyst.

Efficient and practical impregnation methods using vanadium and nickel naphthenate solutions are very different from cyclic deposition of metals onto catalysts in an FCCU. At constant soda level, the mechanism of catalyst deactivation by vanadium poisoning is concerned with the reaction of vanadium with the zeolite aluminum.

As increasing amounts of zeolite alumina are attacked by vanadium and aluminum vanadates are formed. The zeolite surface area and FCC catalyst activity will progressively decrease. The effect of increasing the level of vanadium on the zeolite surface area (ZSA) and kinetic microactivity, k(Ma), of Resoc catalysts is presented in Fig. 5.

The laboratory metals deactivation procedure involved an incipient wetness impregnation of Resoc catalysts with a stock solution of vanadium and nickel naphthenates (2V + Ni) in naphtha. After careful drying, the catalysts were calcined in flowing air at 700 C. for 3 hr.

Catalysts were then steamed at 760 C. in 100% steam at 5 psig for 6 hr. Microactivity testing (MAT) of catalysts loaded with up to 12,000 ppm (2V + Ni) was carried out using a sour, heavy gas oil feed and a cat-to-oil of 4.12.

Zeolite surface areas were taken as the microporous surface area measured by nitrogen adsorption.

As shown in Fig. 5, a stable relationship between ZSA/k(Ma) and k(Ma) exists until the kinetic microactivity decreases from more than 3 down to about 1.38 (Ma = 58). This corresponds to about 6,000 ppm (2V + Ni) on catalyst.

The slope of the line to k(Ma) = 1.38 shows that as the catalyst metals level increases, more zeolite surface area is retained per unit of conversion. As the vanadium level on catalyst increases and k(Ma) decreases to less than about 1.38, the ZSA/k(Ma) vs. k(Ma) relationship becomes more variable.

In the meta-stable region denoted in Fig. 5, the zeolite surface area begins to decrease proportionally greater than the microactivity. The loss in zeolite surface area becomes catastrophic at very high metals levels of about 12,000 ppm (2V+Ni).

The circulating catalyst inventory of an FCCU normally has a microactivity greater than 58. Therefore, the ZSA/k(Ma) vs. k(Ma) relationship for k(Ma) 1.38 should provide a useful tool to monitor the effects of vanadium deactivation in an FCCU.

However, for this to be possible, the same ZSA/k(Ma) vs. k(Ma) relationship must be observed for equilibrium catalysts.

Ten Resoc equilibrium catalysts were chosen at random from nine different European FCCUS. The data presented in Fig. 6 show that the relationship between ZSA/k(Ma) and k(Ma) for the equilibrium catalysts and laboratory catalysts are similar.

This strongly indicates that the mechanism of deactivation of the catalyst after the laboratory impregnation and steaming described previously is similar to that in an actual FCCU.

Individual refineries are identified in Table 1 which presents FCCU specific data for metals on catalyst, average catalyst age, and regenerator temperature. By comparing the data of Table 1 with the activity data presented in Fig. 6, it can be seen that a simple correlation between equilibrium catalyst (Ecat) metals levels and position on the ZSA/k(Ma) vs. k(Ma) plot does not exist. It is perhaps not surprising that additional factors specific to FCCU design do not affect the deactivation characteristics of the catalyst.

For example, steam partial pressure in the regenerator, regenerator combustion mode, and stripping steam are additional factors which influence catalyst activity.

The most desirable ZSA/k(Ma) scenario for a vanadium-tolerant cracking catalyst is when the ZSA/k(Ma) ratio remains approximately constant as the vanadium on catalyst increases. This is shown schematically in Fig. 7.

If the zeolite structure is excessively stable, when vanadium reacts with zeolite aluminum, the catalyst activity decreases and a pure silica zeolite structure remains. For this catalyst design, the ZSA/k(Ma) value will increase as the vanadium level on catalyst increases (Fig. 7).

The pure silica zeolite structure has an extremely low cracking activity but has a high surface area. In an FCCU such catalysts lead to a high E-cat surface area, low E-cat activity, and the FCCU will suffer from a poor stripper operation which will be made worse by the addition of fresh catalyst.

If the FCC catalyst contains a matrix which is more stable than the zeolite (the matrices are calcined at extreme temperatures prior to catalyst formulation) the zeolite surface area will decline more rapidly than the catalyst activity as the vanadium level on the catalyst increases.

For such catalysts, the ZSA/k(Ma) value will be lower at higher catalyst vanadium levels as depicted in Fig. 7. Selective zeolite cracking is replaced by unselective matrix cracking.

In an FCCU as the vanadium level on the E-cat increases, matrix cracking dominates to give high gas and coke yields leading to operating constraints on the gas compressor and on coke burning capacity. Therefore, the design of a vanadium tolerant FCC catalyst must take into account the relative activities of the zeolite and matrix under realistic vanadium deactivation conditions.

IMPROVED ACTIVITY RETENTION

A marked improvement in catalyst vanadium tolerance is achieved by combining high-stability, ultra-stable zeolites with a highly accessible vanadium trap present in the catalyst. Laboratory metals tolerance data for Super Resoc-RV are presented in Fig. 8.

The vertical displacement of the Super Resoc-RV curve to higher ZSA/k(Ma) values (Fig. 7) shows that for a given 2V+Ni loading, more active zeolite surface is available to convert the gas oil feed compared with Resoc catalysts.

FCC conversions of residual oil will benefit from a higher zeolite activity to direct the cracking reactions towards valuable liquid products and away from dry gas and coke.

GASOLINE OCTANES

European refineries are generally demanding two key types of octane catalysts which give highest motor octane and highest motor octane barrels. Because the lead phaseout in Europe is well under way, some refineries are having difficulties in reaching the minimum motor octane specifications for premium (super plus) gasolines.'o "

The clear motor octane number of FCC gasoline can be increased by making more aromatics, by making highly branched light paraffins, and by increasing the olefin-to-paraffin ratio of the light gasoline. 12 13 A modified ultrastable Y zeolite, such as Z-14G, enhances the formation of highly branched light isoparaffins at the expense of olefins.

This hydrocarbon transformation reaction is promoted because the acid sites in Z-14G are more accessible compared with those in Z-14US.

The unit cell sizes of equilibrium catalysts, such as Octacat-D (Z-14G) and DG (Z-14US), are the same at 24.24 A. Thus, when both are steamed, they have the same numbers of Broensted acid sites associated with the aluminum atoms in the zeolite framework.

The yields of light paraffins (Fig. 9) are higher over Z-14G catalysts. A high degree of branching (DOB) of the light paraffins formed over Octacat-DG is seen in the C6 fraction for which the DOB is 12.4 compared to 11.0 for the C6 fraction formed over Octacat-D.

This change in gasoline molecular weight distribution is one reason for an observed increase in the motor octane number of the light gasoline formed over Octacat-DG.

The number of acid sites in rare-earth-exchanged Z-14G has an equilibrium unit cell size of 24.27 A which is the same as for Super Resoc that contains Z-14US. The high acid site density in the Resoc-G zeolite promotes classical hydrogen transfer reactions which convert naphthenes to aromatics as observed for rare-earth-exchanged catalysts.

Thus, Resoc-G gasoline contains a high amount of naphthenic and aromatic ring structures which can be seen from the gasoline composition given in Table 2.

The utilization of olefins as intermediates causes the research octane (RON) of gasoline formed over the "G" catalysts to be equal to, or slightly lower than, that of gasoline formed over FCC catalysts containing conventional zeolites. Therefore, gasoline from the "G" FCC catalysts typically has a low octane sensitivity. 14

Octane numbers calculated for the MAT gasolines using the W.R. Grace octane model are presented in Table 2. The model takes into consideration the blending characteristics of the approximately 300 identified hydrocarbons in the MAT gasoline. 1 3

The "G" FCC catalysts produce gasolines having higher motor octane and lower gasoline sensitivity compared with conventional catalysts at the same unit cell size.

High gasoline yields are obtained over the "G" zeolites because of the increased number of acid sites.

The MAT yield data presented in Table 2 show that the increase in gasoline is obtained at the expense Of C3 and C4 olefins.

For those refineries with alkylation and/or MTBE units, the combination of Resoc-G and Additive O-HS to give good yields of high-octane gasoline and C3 and C4 olefins, especially isobutene, is a profitable and economic catalyst proposal.

The effect of FCCU inventory changeout from Resoc-1 to Resoc-G on the clear MON of the light gasoline produced in a central European refinery is presented in Fig. 10. During the changeout to Resoc-G, the reactor temperature was 525 C., the gasoline yield (C5 - 213 C. EP) was 52.7 0.9 wt %, and the light gasoline clear RON was 93.0 - 0.4. The catalyst change to Resoc-G increased total gasoline octane by more than 1 MON clear at constant gasoline yield.

REFERENCES

Meisenheimer, R.G., Journal of Catalysis, Vol. 1, 1962, pp. 356-362.
Occelli, M.L., Kowalczyk, D.C., and Kibby, C.L., Applied Catalysis, Vol. 16, 1985, pp. 227-236.
Occelli, M.L., Psaras, D., Suib, S.L., and Stencel, J.M., Applied Catalysis, Vol. 28,1986, pp. 143-160.
Occelli, M.L., and Stencel, J.M., in "Fluid Catalytic Cracking- Role in Modern Refining," ACS Symposium Series, Vol. 375, 1987, pp. 195-214.
Tatterson, D.F., and Meiville, R.L., Industrial Engineering Chemistry, Research, Vol. 27, 1988, pp. 1595-1599.
Schubert, P.F., and Altomare, C.A., in "Fluid Catalytic Cracking-Role in Modern Refining," ACS Symposium Series, Vol. 375, 1987, pp. 182-194.
Kugler, E.L., and Leta, D.P., Jounal of Catalysis, Vol. 109, 1988, pp. 387-395.
Wormsbecher, R.F., Peters, A.W., and Maselli, J.M., Journal of Catalysis, Vol. 100, 1986, pp. 130-137.
Maher, P.K., and McDaniel, C.V., U.S. Patent 3 449 070, 1966.
Cartlidge, S., McElhiney, G., and Thompson, J.M., Oil-Gas European magazine, Vol. 2, 1989, p. 26.
Todd, M., European Chemical News, European News Supplement, December 1988, p. 24.
Marcilly, C., and Bourgogne, M., Symposium on Hydrocarbon Chemistry-FCC Naphtha Formation, Division of Petroleum Chemistry, American Chemical Society, Miami, September 1989 p. 763.
Plumlee, K.W., and Cotterman, Symposium on Hydrocarbon Chemistry-FCC Naphtha Formation, Division of Petroleum Chemistry, American Chemical Society, Miami, September 1989 p. 756.
Deady, J., Young, G.W., and Wear, C.C., Paper No. AM-89-13, National Petroleum Refiners Association annual meeting, San Francisco, March 1989.